Upgrading FPGA Implementation of Isolated Word Recognition System for a Real-Time Operation

Abstract

The article reports on the upgrading of the FPGA based isolated word recognition system for real-time tasks. All recognition system components (except some feature calculation steps) were implemented using VHDL. Some high precision calculations were implemented on soft core processor. The employed Dynamic time warping algorithm was speeded-up 2.8 times by restricting the calculated error matrix size. This enabled us to reduce the average word recognition time to 12.81 ms. Linear predictive coding, linear predictive coding cepstral and linear frequency cepstral coefficients feature analyses were investigated for 100 Lithuanian word recognition. In speaker dependent experiments linear predictive coding cepstral analysis gave the highest average recognition rate of 95 % and the highest robustness to white noise in speech. 15 dB noise level lowered average recognition rate to 86.2 %. Index Terms-Cepstral analysis, dynamic time warping, field programmable gate array, intellectual property core, isolated word recognition, linear predictive coefficients.

Wireless Sensor Networks: Towards Resilience Against Weather-Based Disruptions

“…Previously it was considered that meteorological conditions did not affect frequencies below 10 GHz, however, taking into account the research data for the last years, it became obvious that factors such as rain, cold, heat, sunlight and other natural factors can significantly change the performance of wireless networks that operate at frequencies up to 2.4 GHz. Research of network parameters from the point of view of telematics can be expedient, as it can allow real-time optimization of the performance of wireless networks [30], thus improving the quality of services received by the end user. The results of such studies can help network developers and researchers to create new protocolsadapted to the climatic characteristics and weather conditions, which will ensure the best performance of wireless networks and optimize their work.…”

Implementation of a Speech Recognition System in a DSC

“…La mayoría de las soluciones propuestas en la literatura que A. G. Alvarez, Estudiante de la Universidad Tecnológica Nacional, Facultad Regional Buenos Aires, Argentina, alegasalv@gmail.com D. A. Evin, Investigador del Laboratorio de Investigaciones Sensoriales, INIGEM, CONIET-UBA, Buenos Aires, Argentina, diegoevin@gmail.com S. Verrastro, Profesor de la Universidad Tecnológica Nacional, Facultad Regional Buenos Aires, Argentina, sebastian_verrastro@yahoo.com.ar abordan el problema de Reconocimiento Automático del Habla (RAH) en Sistemas Embebidos, se basan en Procesadores de Señales Digitales (DSP) [2][3][4][5][6] o en Field Programmable Gate Arrays (FPGA) [7][8][9]. También pueden encontrarse algunas implementaciones basadas en Microcontroladores.…”

FPGA Design of Voice Enabled Ignition using G279 for Modal Based Speech Compression

“…Some other considerations suggested in the literature are the power consumption, the error rate and error rejection and the speed of processing and memory [9]. Additionally, in voice recognition, the rejection of noise plays an important part especially when environmental variables are beyond control [8] An ignition system, however, allows enough control of the environment to allow filtering stages to remove noise and the speed needs only to be fast enough that the delay does not impact the user's patience.…”